Assessing the quality of biclusters using fuzzy biclustering index
by Nishchal Kumar Verma; Esha Dutta; Yan Cui
International Journal of Data Mining and Bioinformatics (IJDMB), Vol. 15, No. 4, 2016

Abstract: Several algorithms are proposed in the literature for extracting local patterns from a large data matrix. This technique of data mining is known as biclustering. Each of the biclustering algorithms is specialised in extracting different kinds of biclusters. Some algorithms detect equal biclusters, whereas some identify scaled biclusters (Madeira et al., 2004). For any practical database, since we are not aware of the biclusters present in it, we are not sure of the biclustering algorithm to be used. In such a scenario, it is important to define metrics to compare the quality of the extracted biclusters and hence the quality of the biclustering algorithm. In this paper, we have defined novel measures of Hausdorff distance between biclusters and global silhouette index for estimating the quality of biclusters extracted by the existing algorithms. We have also combined these metrics with the proportion of enriched biclusters extracted and defined an overall index defined as the Fuzzy Biclustering Index (FBI) to compare the various algorithms. For a given data set, higher is the FBI, better is the biclustering algorithm.

Online publication date: Thu, 04-Aug-2016

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Data Mining and Bioinformatics (IJDMB):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com